中文摘要：

知识溢出是高科技产业集群中一个突出现象，本文从企业衍生的视角，对知识溢出、知识流失、企业研发以及知识获取等要素进行动态的模型分析，并运用动态面板数据模型，利用系统广义矩估计方法实证分析高科技产业集群中研发与知识溢出效应及其差异性，研究表明，虽然研发对产业集群创新能力存在显著的区域差异性和时滞性，会出现知识溢出的短期效应，即企业知识流失，但是创新知识具有累积效应，随着企业加大研发与学习投入，将会出现知识溢出的长期效应即产业集群中知识存量增加，每个企业都有机会获得和利用更多的外部知识，形成产业集群技术升级的良性循环。

英文摘要：

Knowledge spillover is a prominent phenomenon of hi - tech industrial clusters. From the perspective of spin - off companies, a data model, which mainly includes key elements, such as knowledge spillover, knowledge loss, R＆D, and knowl- edge acquisition, is built to carry out a dynamic analysis. Based on the dynamic panel data model, an empirical analysis is con- ducted to analyze the effect and difference between R＆D and knowledge spillover by using the estimation method of System GMM. The msnlt of the analysis shows that although R＆D exerts a remarkable regional difference and time delay effect on the innovation ability of industrial cluster, and a negative effect of knowledge spillovers i.e. corporate knowledge loss will appear in the hi - tech industry cluster within the short term; if enterprises continue to increase their investment in R＆D, the accumulative effect of knowledge spillover will arise, there would be a long term effect of knowledge spillover, that is, the total amount of knowledge will increase inside the cluster, and each enterprise is able to have the opportunity to get and utilize more external knowledge, thus a virtuous circle of knowledge generation and upgrade will be formed in the hi - teeh industrial cluster.